Have you ever spoken with a translation agency or translator and had the terms “new words”, “fuzzy or partially matched words” and “repeated words” used without them being defined for you?

AZ World has answered this question many times for our clients and potential clients; we thought that it would be useful to share our answer below, as well as describe how these terms could be used by language service providers.

Once we receive documents to translate, AZ World reviews and analyzes the source text (i.e. the text in the originating language) to be translated into “new words”, “fuzzy matched words” and “repeated words” using our computer assisted translation (“CAT”) software – SDL Studio

When AZ World’s team translates a file,we will take less time if we’ve translated a similar segment of text (i.e. a sentence or phrase) previously and it has been saved in our client’s translation memory (“TM”) database.

Prior to starting the translation, we run an analysis of the documents to be translated in SDL Studio and use the following definitions in calculating our Fee Estimate:

  • New words = 0 to 74% matches identified in the SDL Studio Analysis Report
  • Fuzzy matched words = 75 to 99% matches identified in the SDL Studio Analysis Report
  • Repeated words = 99% matches identified in the SDL Studio Analysis Report

We provide our clients with the SDL Studio Analysis Report which provides the number of new, fuzzy match and repeated words as support for our Fee Estimate. If the new translation project is for a client for whom we have previously performed translations, and have created a TM (Translation Memory), we give our client a discount for “fuzzy matched words” and “repeated words” to reflect the actual time and effort spent on that translation project by our translators and editors/proof-readers.

What exactly is a “fuzzy match” or a “partial translation”?

The terms “new words” and “repeated words” are relatively easy for people to understand without a long definition. However, a “fuzzy match” or a “partial translation” is a match in the client’s TM (Translation Memory) identified by the translator’s CAT software (i.e. SDL Studio in AZ World’s case) that ranges from 75-99% matches of the segment of text (i.e. the sentence or phrase) in question in the source text.

For instance, a sentence or phrase with ten words having just one word difference from a sentence already stored in the TM database will result in a 90% fuzzy match; whereas if it has only five words in common with that stored phrase or sentence, then this is a 50% match and this sentence or phrase would be considered a new word. Repeated words are exact matches to a sentence or phrase in the TM.

A “fuzzy match” or a “partial translation” is a segment of text (i.e. a sentence or phrase) in the source text that is similar to a segment that is stored in the TM. It’s only a partial match so it will require some editing. The amount of editing needed is reflected in the percentage. A 60% match will probably need a larger / more complete rewriting, whereas a 99% match may only need to have a comma added, a difference in capitalization or one word changed.

What isn’t a fuzzy match or a partial translation?

A fuzzy match or a partial translation isn’t how many times a common word is repeated in a file. Matches are calculated on segment content – usually a whole sentence or a phrase – and only when a CAT tool is being used

If you are looking to translate something and need some help, please contact us at info@a-zworld.ca or visit us at www.a-zworld.ca.